Today I stumbeld across an entertaining as well as inspiring talk held by David McCandless who showcased some of his data visualizations. He conceives visualizations in general as a “fertile medium”, for a simple reason: It incorporates natural language and vision making complex issues easy to understand, and facilitates the acknowledgement of hidden patterns.

Afterwards, I tried to come up with an idea that would help me understand the topic in more detail – exploring the possibilities, gaining technical insights, revealing critical points. In other words, I was searching for a narrow idea for a simple visualization project that could involve some learning by doing. Furthermore, I wanted to mix the task with linguistics and literature. Later on, my idea arose in a simple question: What about visualizing the main concepts of different literary epochs? Admittedly, that was neither simple nor concrete and apparently esoteric. Hence, I had to dive in some more deliberations.

Some time ago, I took a course in Literature that focused on the epoch of the German Romanticism. Generally, literary epochs cannot be distinguished solely by a timeframe. Some epochs, for example the German Romanticism and Classicism are partly overlapping – in other words, they happened at the same time. More appropriate, we can consider them as a set of literary texts that share a certain set of qualities – in other words: common characteristics. Scientists explore those charactersistics in their research projects. Think of the characteristics of your cookbook. What is unique about its text? Does it cover only vegetarian food? Are there pictures in it? Does it include many lists? To which extend does it differ from your travel guide? Did it affect your life, or did it even affect society? Some cookbooks might. Of course, literary scholars ask and research such questions with respect to literary texts and, of course, way more elaborated. However, where would you begin your search for certain characteristics? In the text, right?

Let us stick with the text – and the words. It is important for me to keep it simple. There is no reason for choking in complexity within a first learning experience. Even though some linguist and literary scholars would doubt the integrity of what I am going to propose, I propose it for simplicity’s sake. I assume that single words bear certain cognitive concepts, pictures or settings that we have in mind when we think about things in life. Hence, words transmit meanings and therefore tell us stories about how the world can be conceived. Think of a good friend writing an email to you. If he uses a lot of words telling you that he is some kind of sad, feeling depressed, can’t get no sleep; you would not conclude that he or she is feeling happy right now.

This is similar to my approach: Asking texts for the prevaling mood they cover. I would like to automatically extract words that represent the main thoughts of an epoch. Afterwards, I want to visualize those words and their importance in a illustrative manner that displays the framework of thinking back then. If I stick to simplicity, this could be done in a tag cloud. More elaborated, the words could grow like plants. In fact, that would contextualise one of the predominant metaphors in German Romanticism: nature, plants, organic growth. Presumably, this would be an advanced topic.

Finally, I have to conduct some technical researches to determine how this task can be done. The texts should be obtained from web-resources, afterwards processed by a script (probably written in Python), and finally visualized using Processing. Of course, there occur several subtasks such as stopword-filtering et cetera. However, I am looking forward to see how all of this can be done.